Marketing with AI Tutors: A Student Project Using Gemini to Build a Campaign

Stop wasting time piecing together courses—build a real campaign with an AI tutor

Students and instructors: you know the pain. Fragmented resources, scattered feedback, and project briefs that ask for creative campaigns but give no way to test them quickly. This student project brief uses Gemini Guided Learning as an AI tutor so teams can design, launch, measure, and iterate a digital launch campaign in a single semester. It’s a learning-by-doing approach that turns classroom theory into measurable skills.

Why this matters in 2026

Late 2025 and early 2026 brought major improvements in AI tutoring and multimodal guidance—tools like Gemini Guided Learning now deliver stepwise curricula, auto-generated experiments, and in-context creative feedback. That means marketing students can simulate paid media, draft dynamic creative assets, and run realistic A/B tests faster than ever.

Employers expect marketers who can blend strategic thinking with quick data-driven iteration. This brief teaches that through project-based learning, while preserving instructor oversight and academic rigor.

Project overview: Marketing with AI Tutors (8–10 week student project)

Teams of 3–4 will use Gemini Guided Learning as their primary AI tutor to create a launch campaign for a real or fictional product. The project emphasizes rapid ideation, experiments, and documented iteration.

Learning goals

• Design a customer-focused strategy using segmentation, value propositions, and buying funnels.
• Use AI tools (Gemini) to generate, test, and refine creative while maintaining critical human judgment.
• Measure campaign performance with real KPIs and apply learnings to iterative improvements.
• Communicate outcomes with a professional launch report and presentation.

Project timeline & deliverables (recommended 8–10 weeks)

1. Week 1 — Project kickoff & discovery
   • Deliverable: One-page creative brief (team)
   • Action: Use Gemini to map user personas and pain points; collect primary/secondary research.
2. Week 2 — Strategy & messaging
   • Deliverable: Messaging hierarchy and positioning statement
   • Action: Ask Gemini for headline variations, benefit statements, proof points, and a 3-tier value ladder.
3. Weeks 3–4 — Creative production
   • Deliverable: 3 hero ads (static or short video), 6 social variations, landing page copy
   • Action: Use Gemini’s asset prompts, iterate on copy + imagery suggestions, export creative specs.
4. Week 5 — Experiment plan
   • Deliverable: A/B test plan with hypothesis, sample sizes, metrics, and timeline
   • Action: Have Gemini calculate reachable sample sizes and expected confidence intervals based on target channels.
5. Weeks 6–7 — Launch & measurement
   • Deliverable: Raw campaign data export and weekly performance dashboard
   • Action: Run experiments (campus ads, small paid budget, or simulated cohorts) and use Gemini to interpret results.
6. Week 8 — Iteration
   • Deliverable: Revised creative and experiment #2 plan
   • Action: Apply learnings and instruct Gemini to produce improved variants targeted to winning segments.
7. Week 9–10 — Final report & presentation
   • Deliverable: Comprehensive launch report (10–12 pages) and 10–12 minute presentation
   • Action: Use Gemini to help draft executive summary, metrics visualization suggestions, and recommendations.

How to use Gemini Guided Learning in each phase

Gemini acts as a structured tutor and a practical co-creator. Below are concrete prompts and workflows tailored to classroom use.

Phase 1 — Research & persona building

Prompt example: "Help our team build two detailed user personas for a low-cost eco-friendly water bottle sold to college students. Include goals, frustrations, channels, and 3 messaging hooks per persona."

Use Gemini to: synthesize literature, create empathy maps, and suggest survey questions for primary research. The AI’s Guided Learning flows can suggest quick micro-assignments (e.g., 30-minute campus intercept surveys) and generate templates for consent and data logging.

Phase 2 — Messaging, copywriting & creative specs

Prompt example: "Give 6 headline-test variants (short, benefit, curiosity, social proof, fear-of-missing-out, and testimonial) plus 3 subheadlines for paid social ads targeting Persona A."

Use the AI to create copy variations, then craft quick heuristics to rank them: clarity, relevance, novelty, and brand fit. Export specs (image size, CTA text, caption) that match common ad platforms.

Phase 3 — Experiment design & simulation

Prompt example: "Design an A/B test to compare Benefit-driven vs. FOMO-driven messaging over 7 days with a $200 ad spend and expected CTR 0.7%. Provide sample sizes and expected time-to-signal."

Gemini can compute sample sizes, suggest statistical significance thresholds, and even simulate outcomes to show likely ranges of conversion metrics. That helps teams avoid underpowered tests and ties directly into a marketer’s playbook for account-level placements and exclusions.

Phase 4 — Analysis & iteration

Prompt example: "We ran two versions: Version A CTR 0.8% CVR 3.2%; Version B CTR 0.5% CVR 4.0%. Recommend next steps and a revised creative hypothesis."

Use Gemini to translate raw metrics into action: identify segment-level winners, propose creative swaps, and generate follow-up experiments that focus on lift and cost-efficiency. For structured post-mortems and instructor feedback loops, refer to micro-feedback workflows to streamline reviews.

Sample prompts and cheat-sheet (worked examples)

Below are compact templates students should copy and adapt. Treat them as a starting point and always add context about the product and target audience.

• Persona prompt: "Create Persona X for [product], age range, motivations, channels, and 3 objections with short rebuttals."
• Headline bank: "Generate 12 short, test-ready headlines using: benefit, curiosity, problem, social proof, scarcity, and question formats."
• Ad variation: "Produce 3 caption lengths (short, medium, long) and 2 CTAs for a Facebook/IG ad."
• Test plan: "Write an A/B test plan: hypothesis, KPI, sample size calc, significance target, stop rules."
• Post-mortem: "Summarize what worked, why, and top 3 recommendations for next campaign."

Metrics students must track (and why)

Good campaigns track both acquisition efficiency and quality of engagement. Make sure your team measures:

• CTR (click-through rate) — early signal of creative relevance.
• CVR (conversion rate) — landing page and funnel health.
• CPA (cost per acquisition) and CPL (cost per lead) — budget efficiency.
• ROAS or LTV-to-CPA — long-term sustainability for paid channels.
• Retention/engagement — downstream quality of acquired users.

Gemini can help calculate these, flag anomalies, and recommend next steps. For classroom fairness, require that teams submit raw data CSVs alongside AI summaries.

Iteration cycles: Build — Measure — Learn

Teach students to run short cycles. Each cycle should have a clear hypothesis, a success metric, and a predetermined analysis window. A common cadence looks like this:

1. Build (1 week) — produce 2–3 creative variants
2. Measure (1–2 weeks) — run tests with minimum sample size
3. Learn (2–3 days) — review results, document insights
4. Pivot or scale — iterate winning variant or test a new variable

Using Gemini speeds up the "Build" step and sharpens hypotheses with suggested micro-experiments derived from earlier results.

Grading rubric — clear, fair, and aligned with learning goals

This rubric is instructor-ready and can be adapted by weight. Total: 100 points.

1. Strategy & research (20 pts) — Quality of personas, market insight, and alignment to objectives (18–20 excellent; 15–17 good; 12–14 acceptable; <12 needs work).
2. Creative execution (20 pts) — Clarity, brand fit, and range of tested creative (20 excellent; 14–19 good; 9–13 acceptable; <9 needs work).
3. Experiment design & data (20 pts) — Statistically sound tests, documented methodology, and raw data submission (20 excellent; 15–19 good; 10–14 acceptable; <10 needs work).
4. Iteration & learning (20 pts) — Evidence-based revisions, rationale for changes, and measurable improvement (20 excellent; 15–19 good; 10–14 acceptable; <10 needs work).
5. Report & presentation (15 pts) — Professionalism, storytelling, and actionable recommendations (15 excellent; 11–14 good; 7–10 acceptable; <7 needs work).
6. Ethics & AI attribution (5 pts) — Proper disclosure of AI use, privacy care, and academic honesty (5 excellent; 3–4 acceptable; <3 needs work).

Notes for instructors: require a short appendix listing all Gemini prompts used and indicate which outputs were directly adopted. This increases transparency and supports grading.

Academic honesty, attribution & ethics

Students must declare AI assistance. A recommended statement in the appendix:

"AI-assisted components: [list prompts and outputs]. Team members verified and adapted all AI outputs. No protected data was used without consent."

Also address data privacy: if running experiments that collect personal data, teams must follow IRB-like consent procedures, redact personal identifiers, and store data securely. For institutional concerns about model auditing and deployment, review best practices for running large language models on compliant infrastructure.

Instructor tips and common pitfalls

• Don’t let AI replace judgment. Use Gemini for speed, not for final decisions.
• Watch for overfitting to small datasets—encourage cross-validation of results.
• Set strict stop rules for experiments (minimum sample, time window) to avoid misreading noise as signal.
• Require raw data exports and a brief walkthrough video of how teams used Gemini to produce key deliverables.

Real-world example (mini case)

Team “River” chose a fictional study app targeted at freshmen. They used Gemini to generate three persona clusters, tested two headline approaches, and ran a campus paid social pilot with a $150 budget. Gemini helped estimate sample sizes and suggested a second-wave CRO (conversion rate optimization) test. The final result: a 35% lift in CTR after swapping to benefit-driven creative and a 12% improvement in landing page CVR after following Gemini’s UX copy recommendations. The team's grade: 92/100—main weakness was incomplete documentation of consent for a campus intercept survey.

This example shows how AI can accelerate learning while still requiring human oversight for ethics and rigor.

Future trends — what students should watch for

• More integrated multimodal guidance: expect AI tutors to suggest complete asset libraries (copy + imagery + cadence).
• Automated experiment orchestration: by 2027, platforms will likely let you schedule, run, and analyze low-cost experiments directly through AI-guided UIs.
• Greater emphasis on privacy-preserving testing: synthetic cohorts and federated test designs will become common.

Actionable takeaways

• Start small: run short, well-powered tests before scaling creative or budget.
• Document everything: keep prompts, raw data, and rationale in an appendix.
• Focus on learning: the grade favors evidence-backed iteration over flashy creative alone.
• Disclose AI use: it’s required for trust and academic integrity.

Submission checklist (for students)

• One-page creative brief
• Persona & research appendix
• All creative assets (source files + specs)
• Raw data CSV(s) from experiments
• Test plans and analysis notebooks
• Final report (10–12 pages) and slide deck
• Appendix with Gemini prompts and AI output citations

Closing — why this brief works

This project brief bridges classroom theory and 2026 workplace expectations by putting AI-powered experimentation into students' hands. It teaches critical thinking, responsible AI use, and the iterative discipline modern marketers need. Gemini Guided Learning shortens the feedback loop, but the real learning comes from how students interpret results and apply them.

Call to action

Ready to run this in your class? Download the printable brief, editable rubric, and prompt cheat-sheet from our community hub at asking.space, try the first two-week sprint with Gemini Guided Learning, and share your results. Post your final deliverable and tag #AIMarketingProject—get peer feedback and optional expert reviews from our network of instructors and practitioners. For inspiration on turning a live launch into a visual story, see this live-launch micro-documentary case study.

Related Reading

• High‑Conversion Product Pages with Composer in 2026: Live Commerce, Scheduling, and Zero‑Trust Workflows
• A Marketer’s Guide to Using Account-Level Placement Exclusions and Negative Keywords Together
• Vertical Video Rubric for Assessment: What Teachers Should Grade in 60 Seconds
• Edge‑First Creator Commerce: Advanced Marketplace Strategies for Indie Sellers in 2026
• Lighting & Optics for Product Photography in Showrooms: 2026 Equipment Guide
• How Vice Media’s C-Suite Shakeup Signals New Opportunities for Production-Focused Creators
• Host a Tech-Savvy Seafood Supper: CES Gadgets That Make Dinner Easy and Memorable
• Gamify Your Home Mobility Routine: Using Level Design Principles to Build Consistency
• Mega Ski Passes: Are They Worth It for Families and Weekend Warriors?
• How AI’s Chip Appetite Is Driving IT Budgeting: A Compatibility-Focused Procurement Playbook